Recently a supplier quality manager asked me about a dilemma he was having with the way his manufacturing facility was measuring in-process supplier quality. If they found defects in supplier parts during the manufacturing process, each defective item was tallied as part of the total. That is, each defective part counted against the total quality performance rather than each shipment counting as a whole against the total. If one shipment was bad and had 25 of the same defects in it, all 25 defects counted against the supplier’s quality performance rather than just counting it as one defective shipment and essentially just one defect. The problem is that his staff is evaluated, in part, on the basis of supplier performance. One supplier quality incident can appear much more serious than it actually is, and it negatively impacts both the supplier’s scores and the performance reviews of the supply management staff that is responsible for managing that supplier.
The supplier quality manager felt that each quality defect should be counted in the overall quality performance score, as each defect is a problem, even if a number of parts have the same defect (and in many cases, a minor defect, which is a whole other issue). His staff feels that this approach is unfair, as one defect is typically addressed as one problem, even if it has occurred multiple times within a shipment. Mostly, however, they seem to be unhappy about the larger impact on their own performance reviews.
What should this company do? One approach is to calculate quality performance as they do now, but take the frequency and severity of quality incidents into account in the performance review. If there is one incident with 25 parts, the staff reasons, it’s much better than 25 incidents involving one part. And the one larger-size incident should have less impact on the performance review compared to multiple incidents with fewer parts. This approach is more subjective at review time and depends on the discretion of the manager. Is this subjectivity unfair? What do you think of each approach?
I would love to hear your views on how this company should handle this measurement challenge.